Published on

Published in: Technology
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide


  1. 1. Lectures on the Geometry of Quantization Sean Bates Alan Weinstein Department of Mathematics Department of Mathematics Columbia University University of California New York, NY 10027 USA Berkeley, CA 94720 USA 1
  2. 2. Contents1 Introduction: The Harmonic Oscillator 52 The WKB Method 8 2.1 Some Hamilton-Jacobi preliminaries . . . . . . . . . . . . . . . . . . . . . . . 8 2.2 The WKB approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Symplectic Manifolds 17 3.1 Symplectic structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2 Cotangent bundles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.3 Mechanics on manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324 Quantization in Cotangent Bundles 36 4.1 Prequantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4.2 The Maslov correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.3 Phase functions and lagrangian submanifolds . . . . . . . . . . . . . . . . . . 46 4.4 WKB quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565 The Symplectic Category 64 5.1 Symplectic reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 5.2 The symplectic category . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 5.3 Symplectic manifolds and mechanics . . . . . . . . . . . . . . . . . . . . . . 796 Fourier Integral Operators 83 6.1 Compositions of semi-classical states . . . . . . . . . . . . . . . . . . . . . . 83 6.2 WKB quantization and compositions . . . . . . . . . . . . . . . . . . . . . . 877 Geometric Quantization 93 7.1 Prequantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 7.2 Polarizations and the metaplectic correction . . . . . . . . . . . . . . . . . . 98 7.3 Quantization of semi-classical states . . . . . . . . . . . . . . . . . . . . . . . 1098 Algebraic Quantization 111 8.1 Poisson algebras and Poisson manifolds . . . . . . . . . . . . . . . . . . . . . 111 8.2 Deformation quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 8.3 Symplectic groupoids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114A Densities 119B The method of stationary phase 121 ˇC Cech cohomology 124D Principal T bundles 125 2
  3. 3. PrefaceThese notes are based on a course entitled “Symplectic geometry and geometric quantization”taught by Alan Weinstein at the University of California, Berkeley, in the fall semester of1992 and again at the Centre Emile Borel (Institut Henri Poincar´) in the spring semester eof 1994. The only prerequisite for the course (and for these notes) was a knowledge ofthe basic notions from the theory of differentiable manifolds (differential forms, vector fields,transversality, etc.). The aim of the course was to give students an introduction to the ideas ofmicrolocal analysis and the related symplectic geometry, with an emphasis on the role whichthese ideas play in formalizing the transition between the mathematics of classical dynamics(hamiltonian flows on symplectic manifolds) and that of quantum mechanics (unitary flowson Hilbert spaces). There already exist many books on the subjects treated here, but most of them providetoo much detail for the reader who just wants to find out what the subject is about. Thesenotes are meant to function as a guide to the literature; we refer to other sources for manydetails which are omitted here, and which can be bypassed on a first reading. The pamphlet [63] is in some sense a precursor to these notes. On the other hand, a muchmore complete reference on the subject, written at about the same time, is [28]. An earlierwork, one of the first to treat the connections between classical and quantum mechanics froma geometric viewpoint, is [41]. The book [29] treats further topics in symplectic geometryand mechanics, with special attention to the role of symmetry groups, a topic pretty muchignored in the present notes. For more extensive treatment of the PDE aspects of the subject,we refer to [43] for a physics-oriented presentation and to the notes [21] and the treatises[32], [46], and [56]. For “geometric quantization”, one may consult [35], [53], [54], [60] or[71]. For classical mechanics and symplectic geometry, we suggest [1], [2], [6], [8], [25], [38],[59]. Finally, two basic references on quantum mechanics itself are [13] and [20]. Although symplectic geometry is like any field of mathematics in having its definitions,theorems, etc., it is also a special way of looking at a very broad part of mathematics andits applications. For many “symplecticians”, it is almost a religion. A previous paper byone of us [64] referred to “the symplectic creed”.1 In these notes, we show how symplecticgeometry arises from the study of semi-classical solutions to the Schr¨dinger equation, and in oturn provides a geometric foundation for the further analysis of this and other formulationsof quantum mechanics. These notes are still not in final form, but they have already benefitted from the comments 1 We like the following quotation from [4] very much: In recent years, symplectic and contact geometries have encroached on all areas of mathemat- ics. As each skylark must display its comb, so every branch of mathematics must finally display symplectisation. In mathematics there exist operations on different levels: functions acting on numbers, operators acting on functions, functors acting on operators, and so on. Symplec- tisation belongs to the small set of highest level operations, acting not on details (functions, operators, functors), but on all the mathematics at once. Although some such highest level operations are presently known (for example, algebraisation, Bourbakisation, complexification, superisation, symplectisation) there is as yet no axiomatic theory describing them. 3
  4. 4. and suggestions of many readers, especially Maurice Garay, Jim Morehead, and DmitryRoytenberg. We welcome further comments. We would like to thank the Centre Emile Boreland the Isaac Newton Institute for their hospitality. During the preparation of these notes,S.B. was supported by NSF graduate and postdoctoral fellowships in mathematics. A.W.was partially supported by NSF Grants DMS-90-01089 and 93-01089. 4
  5. 5. 1 Introduction: The Harmonic OscillatorIn these notes, we will take a “spiral” approach toward the quantization problem, beginningwith a very concrete example and its proposed solution, and then returning to the samekind of problem at progressively higher levels of generality. Specifically, we will start withthe harmonic oscillator as described classically in the phase plane R2 and work toward theproblem of quantizing arbitrary symplectic manifolds. The latter problem has taken on anew interest in view of recent work by Witten and others in the area of topological quantumfield theory (see for example [7]).The classical pictureThe harmonic oscillator in 1 dimension is described by Newton’s differential equation: m¨ = −kx. xBy a standard procedure, we can convert this second-order ordinary differential equationinto a system of two first-order equations. Introducing the “phase plane” R2 with positionand momentum coordinates (q, p), we set q=x p = mx, ˙so that Newton’s equation becomes the pair of equations: p q= ˙ p = −kq. ˙ mIf we now introduce the hamiltonian function H : R2 → R representing the sum of kineticand potential energies, p2 kq 2 H(q, p) = + 2m 2then we find ∂H ∂H q= ˙ p=− ˙ ∂p ∂qThese simple equations, which can describe a wide variety of classical mechanical systemswith appropriate choices of the function H, are called Hamilton’s equations.2 Hamilton’sequations define a flow on the phase plane representing the time-evolution of the classicalsystem at hand; solution curves in the case of the harmonic oscillator are ellipses centeredat the origin, and points in the phase plane move clockwise around each ellipse. We note two qualitative features of the hamiltonian description of a system: 1. The derivative of H along a solution curve is dH ∂H ∂H = q+ ˙ p = −pq + q p = 0, ˙ ˙˙ ˙˙ dt ∂q ∂p 2 If we had chosen x rather than mx as the second coordinate of our phase plane, we would not have ˙ ˙arrived at this universal form of the equations. 5
  6. 6. i.e., the value of H is constant along integral curves of the hamiltonian vector field. Since H represents the total energy of the system, this property of the flow is interpreted as the law of conservation of energy. 2. The divergence of the hamiltonian vector field XH = (q, p) = ( ∂H , − ∂H ) is ˙ ˙ ∂p ∂q ∂2H ∂2H · XH = − = 0. ∂q ∂p ∂p ∂q Thus the vector field XH is divergence-free, and its flow preserves area in the phase plane.The description of classical hamiltonian mechanics just given is tied to a particular coordinatesystem. We shall see in Chapter 3 that the use of differential forms leads to a coordinate-free description and generalization of the hamiltonian viewpoint in the context of symplecticgeometry.The quantum mechanical pictureIn quantum mechanics, the motion of the harmonic oscillator is described by a complex-valued wave function ψ(x, t) satisfying the 1-dimensional Schr¨dinger equation: o 2 ∂ψ ∂2ψ k 2 i =− + x ψ. ∂t 2m ∂x2 2Here, Planck’s constant has the dimensions of action (energy × time). Interpreting theright hand side of this equation as the result of applying to the wave function ψ the operator 2 ˆ def ∂2 k H = − + m x2 , 2m ∂x2 2where mx2 is the operator of multiplication by x2 , we may rewrite the Schr¨dinger equation oas ∂ψ ˆ i= Hψ. ∂tA solution ψ of this equation does not represent a classical trajectory; instead, if ψ isnormalized, i.e. ψ ∗ ψ = 1, Rthen its square-norm ρ(x, t) = |ψ(x, t)|2is interpreted as a probability density for observing the oscillator at the position x at timet. The wave function ψ(x, t) itself may be viewed alternatively as a t-dependent function ofx, or as a path in the function space C ∞ (R, C). From the latter point of view, Schr¨dinger’s o ∞equation defines a vector field on C (R, C) representing the time evolution of the quantumsystem; a wave function satisfying Schr¨dinger’s equation then corresponds to an integral ocurve of the associated flow. Like Hamilton’s equations in classical mechanics, the Schr¨dinger equation is a general oform for the quantum mechanical description of a large class of systems. 6
  7. 7. Quantization and the classical limitThe central aim of these notes is to give a geometric interpretation of relationships betweenthe fundamental equations of classical and quantum mechanics. Based on the present dis-cussion of the harmonic oscillator, one tenuous connection can be drawn as follows. To theclassical position and momentum observables q, p we associate the differential operators q → q = mx ˆ ∂ p → p = −i ˆ . ∂xThe classical hamiltonian H(q, p) = p2 /2m+kq 2 /2 then corresponds naturally to the operator ˆH. As soon as we wish to “quantize” a more complicated energy function, such as (1 + q 2 )p2 ,we run in to the problem that the operators q and p do not commute with one another, so ˆ ˆthat we are forced to choose between (1 + q )ˆ and p2 (1 + q 2 ), among a number of other 2 2 ˆ p ˆ ˆpossibilities. The difference between these choices turns out to become small when → 0.But how can a constant approach zero? Besides the problem of “quantization of equations,” we will also treat that of “quan-tization of solutions.” That is, we would like to establish that, for systems which are insome sense macroscopic, the classical motions described by solutions of Hamilton’s equa-tions lead to approximate solutions of Schr¨dinger’s equation. Establishing this relation obetween classical and quantum mechanics is important, not only in verifying that the theo-ries are consistent with the fact that we “see” classical behavior in systems which are “really”governed by quantum mechanics, but also as a tool for developing approximate solutions tothe quantum equations of motion. What is the meaning of “macroscopic” in mathematical terms? It turns out that goodapproximate solutions of Schr¨dinger’s equation can be generated from classical information owhen is small enough. But how can a constant with physical dimensions be small? Although there remain some unsettled issues connected with the question, “How can become small?” the answer is essentially the following. For any particular mechanicalsystem, there are usually characteristic distances, masses, velocities, . . . from which a unitof action appropriate to the system can be derived, and the classical limit is applicablewhen divided by this unit is much less than 1. In these notes, we will often regardmathematically as a formal parameter or a variable rather than as a fixed number. 7
  8. 8. 2 The WKB MethodA basic technique for obtaining approximate solutions to the Schr¨dinger equation from oclassical motions is called the WKB method, after Wentzel, Kramers, and Brillouin. (Othernames, including Liouville, Green, and Jeffreys are sometimes attached to this method.References [13] and [47] contain a discussion of some of its history. Also see [5], where themethod is traced all the way back to 1817. For convenience, nevertheless, we will still referto the method as WKB.) A good part of what is now called microlocal analysis can beunderstood as the extension of the basic WKB idea to more precise approximations andmore general situations, so the following discussion is absolutely central to these notes.2.1 Some Hamilton-Jacobi preliminariesIn this section we will carry out the first step in the WKB method to obtain an approximatesolution to the “stationary state” eigenvalue problem arising from the Schr¨dinger equation. oThe geometric interpretation of this technique will lead to a correspondence between classicaland quantum mechanics which goes beyond the one described in Chapter 1. Consider a 1-dimensional system with hamiltonian p2 H(q, p) = + V (q), 2mwhere V (x) is a potential (for example the potential kx2 /2 for the harmonic oscillator).Hamilton’s equations now become p q= ˙ p = −V (q). ˙ mFor fixed ∈ R+ , Schr¨dinger’s equation assumes the form o ∂ψ ˆ i = Hψ, ∂twhere 2 ˆ ∂2 H=− + mV 2m ∂x2is the Schr¨dinger operator. o As a first step toward solving the Schr¨dinger equation, we look for stationary states, oi.e. solutions of the form ψ(x, t) = ϕ(x) e−iωt ,so-called because as time evolves, these solutions keep the same form (up to multiplication bya complex scalar of norm 1). Substituting this expression for ψ in the Schr¨dinger equation, owe obtain ˆ ω ϕ(x) e−iωt = (Hϕ)(x) e−iωt .Eliminating the factor e−iωt above, we arrive at the time-independent Schr¨dinger equa- otion: ˆ (H − E) ϕ = 0, 8
  9. 9. where E = ω. This equation means that ϕ is to be an eigenfunction of the linear differential ˆoperator H; the eigenvalue E represents the energy of the system, which has a “definite value”in this state. Suppose for the moment that the potential V is constant, in which case the force −V (x)is zero, and so we are dealing with a free particle. Trying a solution of the form ϕ(x) = eixξfor some constant ξ, we find that ˆ (H − E) ϕ = 0 ⇔ ( ξ)2 = 2m(E − V ).For V < E, the (real) value of ξ is thus determined up to a choice of sign, and one has anabundance of exact solutions of the Schr¨dinger equation which are oscillatory and bounded. oSuch a wave function is not square-integrable and as such is said to be “unnormalizable”; itrepresents a particle which is equally likely to be anywhere in space, but which has a definitemomentum (since it is an eigenfunction of the momentum operator p).3 When E < V , the ˆconstant ξ is imaginary, and there are only real exponential solutions, which are unboundedand admit no physical interpretation. The basic idea at this stage of the WKB method is that, if V varies with x, then ξ shouldvary with x as well; a more general solution candidate is then ϕ(x) = eiS(x)/ ,for some real-valued function S known as a phase function. This proposed form of thesolution is the simplest version of the WKB ansatz, and in this case we have ˆ (S (x))2 i (H − E) ϕ = + (V − E) − S (x) eiS(x)/ . 2m 2mSince we will consider to be “small”, our first-order approximation attempt will ignore thelast term in brackets; to kill the other two terms, we require that S satisfy the eikonal orHamilton-Jacobi equation: (S (x))2 H(x, S (x)) = + V (x) = E, 2mi.e. S (x) = ± 2m(E − V (x)).To understand the phase function S geometrically, we consider the classical phase4 planeR2 T ∗ R with coordinates (q, p). The differential dS = S dx can be viewed as a mappingdS : R → T ∗ R, where as usual we set p = S . Then S satisfies the Hamilton-Jacobi equationif and only if the image of dS lies in the level manifold H −1 (E). This observation establishesa fundamental link between classical and quantum mechanics: When the image of dS lies in a level manifold of the classical hamiltonian, the function S may be taken as the phase function of a first-order approximate solu- tion of Schr¨dinger’s equation. o 3 See [55] for a group-theoretic interpretation of such states. 4 These two uses of the term “phase” seem to be unrelated! 9
  10. 10. The preceding discussion generalizes easily to higher dimensions. In Rn , the Schr¨dinger ooperator corresponding to the classical hamiltonian p2 i H(q, p) = + V (q) 2mis 2 ˆ H=− ∆ + mV , 2mwhere ∆ denotes the ordinary Laplacian. As before, if we consider a WKB ansatz of theform ϕ = eiS/ , then 2 ˆ S i (H − E) ϕ = + (V − E) − ∆S eiS/ 2m 2mwill be O( ) provided that S satisfies the Hamilton-Jacobi equation: 2 ∂S ∂S S(x) H x1 , . . . , xn , ,... , = + V (x) = E. ∂x1 ∂xn 2m ˆSince ϕ is of order zero in , while (H − E) ϕ = O( ), the ansatz ϕ again constitutes afirst-order approximate solution to the time-independent Schr¨dinger equation. o n We will call a phase function S : R → R admissible if it satisfies the Hamilton-Jacobiequation. The image L = im(dS) of the differential of an admissible phase function S ischaracterized by three geometric properties: 1. L is an n-dimensional submanifold of H −1 (E). 2. The pull-back to L of the form αn = j pj dqj on R2n is exact. 3. The restriction of the canonical projection π : T ∗ Rn → Rn to L induces a diffeomor- phism L Rn . In other words, L is projectable.While many of the basic constructions of microlocal analysis are motivated by operationson these projectable submanifolds of T ∗ Rn R2n , applications of the theory require us toextend the constructions to more general n-dimensional submanifolds of R2n satisfying onlya weakened version of condition (2) above, in which “exact” is replaced by “closed”. Suchsubmanifolds are called lagrangian. For example, the level sets for the 1-dimensional harmonic oscillator are lagrangian sub-manifolds in the phase plane. A regular level curve of the hamiltonian is an ellipse L. SinceL is 1-dimensional, the pull-back to L of the form p dq is closed, but the integral of p dqaround the curve equals the enclosed nonzero area, so its pull-back to L is not exact. It isalso clear that the curve fails to project diffeomorphically onto R. From the classical stand-point, the behavior of an oscillator is nevertheless completely described by its trajectory,suggesting that in general the state of a system should be represented by the submanifoldL (projectable or not) rather than by the phase function S. This idea, which we will clarifylater, is the starting point of the geometrical approach to microlocal analysis. 10
  11. 11. For now, we want to note an important relationship between lagrangian submanifolds ofR2n and hamiltonian flows. Recall that to a function H : R2n → R, Hamilton’s equationsassociate the vector field ∂ ∂ ∂H ∂ ∂H ∂ XH = q ˙ +p ˙ = − . ∂q ∂p j ∂pj ∂qj ∂qj ∂pjA simple computation shows that XH and the form αn are related by the equation XH dαn = −dH,i.e. dαn (XH , v) = −dH(v)for every tangent vector v. If L is a lagrangian submanifold of a level set of H, then T Llies in the kernel of dH at all points of L, or, in other words, the 2-form dαn vanishes onthe subspace of Tp R2n generated by Tp L and XH (p) for each p ∈ L. The restriction of dαnto the tangent space Tp R2n of R2n at any point p defines a nondegenerate, skew-symmetricbilinear form, and thus, as we will see in the next chapter, subspaces of Tp R2n on which dαnvanishes can be at most n-dimensional. These remarks imply that XH is tangent to L, andwe have the following result.Hamilton-Jacobi theorem . A function H : R2n → R is locally constant on a lagrangiansubmanifold L ⊂ R2n if and only if the hamiltonian vector field XH is tangent to L.If the lagrangian submanifold L is locally closed, this theorem implies that L is invariantunder the flow of XH .2.2 The WKB approximationReturning to our WKB ansatz for a stationary-state solution of the Schr¨dinger equation, owe recall that if S : R → R is an admissible phase function, then ϕ(x) = eiS(x)/ satisfies n ˆ (H − E) ϕ = O( ). ˆUp to terms of order , in other words, ϕ is an eigenfunction of H with eigenvalue E. There is no way to improve the order of approximation simply by making a better choiceof S. It is also clear on physical grounds that our ansatz for ϕ is too restrictive becauseit satisfies |ϕ(x)| = 1 for all x. In quantum mechanics, the quantity |ϕ(x)|2 represents theprobability of the particle being at the position x, and there is no reason for this to beconstant; in fact, it is at least intuitively plausible that a particle is more likely to be foundwhere it moves more slowly, i.e., where its potential energy is higher. We may therefore hopeto find a better approximate solution by multiplying ϕ by an “amplitude function” a ϕ(x) = eiS(x)/ a(x). 11
  12. 12. If S is again an admissible phase function, we now obtain: ˆ 1 ∂a ∂S 2 (H − E) ϕ = − i a∆S + 2 + ∆a eiS/ . 2m j ∂xj ∂xj ˆIf a is chosen to kill the coefficient of on the right, then ϕ will be an eigenfunction of Hmodulo terms of order O( 2 ). This condition on a is known as the homogeneous transportequation: ∂a ∂S a∆S + 2 = 0. j ∂xj ∂xjIf S is an admissible phase function, and a is an amplitude which satisfies the homogeneoustransport equation, then the second-order solution ϕ = eiS/ a is called the semi-classicalapproximation.Example 2.1 In 1 dimension, the homogeneous transport equation amounts to aS + 2a S = 0.Solving this equation directly, we obtain a2 S + 2aa S = (a2 S ) = 0 c ⇒ a= √ Sfor some constant c. Since S is assumed to satisfy the Hamilton-Jacobi equation, we haveS = 2m(E − V ), and thus c a= 1 . [4(E − V )] 4If E > V (x) for all x ∈ R, then a is a smooth solution to the homogeneous transport equation.Notice that a = |ϕ| is largest where V is largest, as our physical reasoning predicted. Since the expression above for a does not depend explicitly on the phase function S, wemight naively attempt to use the same formula when im(dS) is replaced by a non-projectablelagrangian submanifold of H −1 (E). Consider, for example, the unbounded potential V (x) = √x2 in the case of the harmonic oscillator. For |x| < E, the function a is still well-defined √up to a multiplicative constant. At |x| = E, however, a has (asymptotic) singularities;observe that these occur precisely at the projected image of those points of L where the √projection itself becomes singular. Outside the interval |x| ≤ E, the function a assumescomplex values. To generate better approximate solutions to the eigenfunction problem, we can extendthe procedure above by adding to the original amplitude a = a0 certain appropriately chosenfunctions of higher order in . Consider the next approximation ϕ = eiS/ (a0 + a1 ). 12
  13. 13. Assuming that eiS/ a0 is a semi-classical approximate solution, we obtain: ˆ 1 2 ∂a1 ∂S 3 (H − E) ϕ = − i a1 ∆S + 2 − i∆a0 + ∆a1 eiS(x)/ . 2m j ∂xj ∂xjEvidently, ϕ will be a solution of the time-independent Schr¨dinger equation modulo terms oof order O( 3 ) provided that a1 satisfies the inhomogeneous transport equation ∂a1 ∂S a1 ∆S + 2 = i∆a0 . j ∂xj ∂xj n In general, a solution to the eigenfunction problem modulo terms of order O( ) is givenby a WKB ansatz of the form ϕ = eiS/ (a0 + a1 + · · · + an n ),where S satisfies the Hamilton-Jacobi equation, a0 satisfies the homogeneous transport equa-tion, and for each k > 0, the function ak satisfies the inhomogeneous transport equation: ∂ak ∂S ak ∆S + 2 = i∆ak−1 . j ∂xj ∂xjThis situation can be described in the terminology of asymptotic series as follows. By an -dependent function f on Rn we will mean a function f : Rn × R+ → C, where isviewed as a parameter ranging in R+ . Such a function is said to be represented by aformal asymptotic expansion of the form ∞ ak k , where each coefficient ak is a smooth k=0complex-valued function on Rn , if, for each K ∈ Z+ , the difference K k f − ak k=0is O( K+1 ) locally uniformly in x. When f admits such an expansion, its coefficients ak areuniquely determined. It is obvious that any -dependent function which extends smoothlyto = 0 is represented by an asymptotic series, and a theorem of E.Borel (see [28, p.28])tells us that, conversely, any asymptotic series can be “summed” to yield such a function.The principal part of an asymptotic series ∞ ak k is defined as its first term which is k=0not identically zero as a function of x. The order of a is the index of its principal part. If we consider as equivalent any two -dependent functions whose difference is O( ∞ ), i.e.O( k ) for all k, then each asymptotic series determines a unique equivalence class. A WKB ˆ“solution” to the eigenfunction problem Hϕ = Eϕ is then an equivalence class of functionsof the form ϕ = eiS/ a,where S is an admissible phase function and a is an -dependent function represented by aformal asymptotic series ∞ k a∼ ak k=p 13
  14. 14. with the property that its principal part ap satisfies the homogeneous transport equation ∂ap ∂S ap ∆S + 2 = 0, j ∂xj ∂xjand for k > p, the ak satisfy the recursive transport equations: ∂ak ∂S ak ∆S + 2 = i∆ak−1 . j ∂xj ∂xjThis means that the -dependent function ϕ (or any function equivalent to it) satisfies theSchr¨dinger equation up to terms of order O( ∞ ). oGeometry of the transport equationIn Section 2.1, we saw that a first-order WKB approximate solution ϕ = eiS/ to the time-independent Schr¨dinger equation depended on the choice of an admissible phase function, oi.e., a function S satisfying the Hamilton-Jacobi equation H(x, ∂S ) = E. The generalized or ∂xgeometric version of such a solution was a lagrangian submanifold of the level set H −1 (E).We now wish to interpret and generalize in a similar way the semi-classical approximationwith its amplitude included. To begin, suppose that a is a function on Rn which satisfies the homogeneous transportequation: ∂a ∂S a∆S + 2 = 0. j ∂xj ∂xjAfter multiplying both sides of this equation by a, we can rewrite it as: ∂ ∂S a2 = 0, j ∂xj ∂xjwhich means that the divergence of the vector field a2 S is zero. Rather than consideringthe transport equation as a condition on the vector field a2 S (on Rn ) per se, we can lift allof this activity to the lagrangian submanifold L = im(dS). Notice first that the restrictionto L of the hamiltonian vector field associated to H(q, p) = p2 /2 + V (q) is i ∂S ∂ ∂V ∂ XH |L = − . j ∂xj ∂qj ∂qj ∂pj (x)The projection XH of XH |L onto Rn (the (x) reminds us of the coordinate x on Rn ) there- (x)fore coincides with S, and so the homogeneous transport equation says that a2 XH isdivergence-free for the canonical density |dx| = |dx1 ∧ · · · ∧ dxn | on Rn . But it is better toreformulate this condition as: LX (x) (a2 |dx|) = 0; H 14
  15. 15. (x)that is, we transfer the factor of a2 from the vector field XH = S to the density |dx|. SinceXH is tangent to L by the Hamilton-Jacobi theorem, and since the Lie derivative is invariantunder diffeomorphism, this equation is satisfied if and only if the pull-back of a2 |dx| to Lvia the projection π is invariant under the flow of XH . This observation, together with the fact that it is the square of a which appears in thedensity π ∗ (a2 |dx|), suggests that a solution of the homogeneous transport equation shouldbe represented geometrically by a half-density on L, invariant by XH . (See Appendix A fora discussion of densities of fractional order.) In other words, a (geometric) semi-classical state should be defined as a lagrangian sub-manifold L of R2n equipped with a half-density a. Such a state is stationary when L lies ina level set of the classical hamiltonian and a is invariant under its flow.Example 2.2 Recall that in the case of the 1-dimensional harmonic oscillator, stationaryclassical states are simply those lagrangian submanifolds of R2 which coincide with theregular level sets of the classical hamiltonian H(q, p) = (p2 + kq 2 )/2. There is a unique (upto a constant) invariant volume element for the hamiltonian flow of H on each level curveH. Any such level curve L, together with a square root of this volume element, constitutesa semi-classical stationary state for the harmonic oscillator.Notice that while a solution to the homogeneous transport equation in the case of the 1-dimensional harmonic oscillator was necessarily singular (see Example 2.1), the semi-classicalstate described in the preceding example is a perfectly smooth object everywhere on thelagrangian submanifold L. The singularities arise only when we try to transfer the half-density from L down to configuration space. Another substantial advantage of the geometricinterpretation of the semi-classical approximation is that the concept of an invariant half-density depends only on the hamiltonian vector field XH and not on the function S, so itmakes sense on any lagrangian submanifold of R2n lying in a level set of H. This discussion leads us to another change of viewpoint, namely that the quantum statesthemselves should be represented, not by functions, but by half-densities on configurationspace Rn , i.e. elements of the intrinsic Hilbert space HRn (see Appendix A). Stationary states ˆare then eigenvectors of the Schr¨dinger operator H, which is defined on the space of smooth ohalf-densities in terms of the old Schr¨dinger operator on functions, which we will denote omomentarily as Hfunˆ , by the equation ˆ ˆ H(a|dx|1/2 ) = (Hfun a)|dx|1/2 .From this new point of view, we can express the result of our analysis as follows: If S is an admissible phase function and a is a half-density on L = im(dS) which is invariant under the flow of the classical hamiltonian, then eiS/ (dS)∗ a is a second-order approximate solution to the time-independent Schr¨dinger equation. o 15
  16. 16. In summary, we have noted the following correspondences between classical and quantummechanics: Object Classical version Quantum version basic space R2n H Rn state lagrangian submanifold of R2n with half-density on Rn half-density time-evolution Hamilton’s equations Schr¨dinger equation o generator of evolution function H on R2n ˆ operator H on smooth half-densities stationary state lagrangian submanifold in level set ˆ eigenvector of H of H with invariant half-density Proceeding further, we could attempt to interpret a solution of the recursive system ofinhomogeneous transport equations on Rn as an asymptotic half-density on L in order toarrive at a geometric picture of a complete WKB solution to the Schr¨dinger equation. This, ohowever, involves some additional difficulties, notably the lack of a geometric interpretationof the inhomogeneous transport equations, which lie beyond the scope of these notes. Instead,we will focus on two aspects of the semi-classical approximation. First, we will extend thegeometric picture presented above to systems with more general phase spaces. This willrequire the concept of symplectic manifold, which is introduced in the following chapter.Second, we will “quantize” semi-classical states in these symplectic manifolds. Specifically,we will try to construct a space of quantum states corresponding to a general classicalphase space. Then we will try to construct asymptotic quantum states corresponding tohalf-densities on lagrangian submanifolds. In particular, we will start with an invarianthalf-density on a (possibly non-projectable) lagrangian submanifold of R2n and attempt touse this data to construct an explicit semi-classical approximate solution to Schr¨dinger’s oequation on Rn . 16
  17. 17. 3 Symplectic ManifoldsIn this chapter, we will introduce the notion of a symplectic structure on a manifold, moti-vated for the most part by the situation in R2n . While some discussion will be devoted tocertain general properties of symplectic manifolds, our main goal at this point is to developthe tools needed to extend the hamiltonian viewpoint to phase spaces associated to generalfinite-dimensional configuration spaces, i.e. to cotangent bundles. More general symplecticmanifolds will reappear as the focus of more sophisticated quantization programs in laterchapters. We refer to [6, 29, 63] for thorough discussions of the topics in this chapter.3.1 Symplectic structuresIn Section 2.1, a lagrangian submanifold of R2n was defined as an n-dimensional submanifoldL ⊂ R2n on which the exterior derivative of the form αn = pi dqi vanishes; to a functionH : R → R, we saw that Hamilton’s equations associate a vector field XH on R2n satisfying 2n XH dαn = −dH.Finally, our proof of the Hamilton-Jacobi theorem relied on the nondegeneracy of the 2-form dαn . These points already indicate the central role played by the form dαn in thestudy of hamiltonian systems in R2n ; the correct generalization of the hamiltonian picture toarbitrary configuration spaces relies similarly on the use of 2-forms with certain additionalproperties. In this section, we first study such forms pointwise, collecting pertinent factsabout nondegenerate, skew-symmetric bilinear forms. We then turn to the definition ofsymplectic manifolds.Linear symplectic structuresSuppose that V is a real, m-dimensional vector space. A bilinear form ω : V × V → R givesrise to a linear map ω: V → V ∗ ˜defined by contraction: ω (x)(y) = ω(x, y). ˜The ω-orthogonal to a subspace W ⊂ V is defined as W ⊥ = {x ∈ V : W ⊂ ker ω (x)}. ˜If ω is an isomorphism, or in other words if V ⊥ = {0}, then the form ω is said to be nonde- ˜generate; if in addition ω is skew-symmetric, then ω is called a linear symplectic structure onV . A linear endomorphism of a symplectic vector space (V, ω) which preserves the form ωis called a linear symplectic transformation, and the group of all such transformationsis denoted by Sp(V ). 17
  18. 18. Example 3.1 If E is any real n-dimensional vector space with dual E ∗ , then a linear sym-plectic structure on V = E ⊕ E ∗ is given by ω((x, λ), (x , λ )) = λ (x) − λ(x ).With respect to a basis {xi } of E and a dual basis {λi } of E ∗ , the form ω is represented bythe matrix 0 I ω= . −I 0It follows that if a linear operator on V is given by the real 2n × 2n matrix A B T = , C Dthen T is symplectic provided that At C, B t D are symmetric, and At D − C t B = I. Note inparticular that these conditions are satisfied if A ∈ GL(E), D = (At )−1 , and B = C = 0,and so GL(E) is isomorphic to a subgroup Gl(E) of Sp(V ). More generally, if K : E → Fis an isomorphism, then the association (x, λ) → (Kx, (K −1 )∗ λ)defines a linear symplectomorphism between E ⊕ E ∗ and F ⊕ F ∗ equipped with these linearsymplectic structures.Since the determinant of a skew-symmetric m × m matrix is zero if m is odd, the existenceof a linear symplectic structure on a vector space V implies that V is necessarily even-dimensional and therefore admits a complex structure, i.e. a linear endomorphism J suchthat J 2 = −I. A complex structure is said to be compatible with a symplectic structure onV if ω(Jx, Jy) = ω(x, y)and ω(x, Jx) > 0for all x, y ∈ V . In other words, J is compatible with ω (we also call it ω-compatible)if J : V → V is a linear symplectomorphism and gJ (·, ·) = ω(·, J·) defines a symmetric,positive-definite bilinear form on V .Theorem 3.2 Every symplectic vector space (V, ω) admits a compatible complex structure.Proof. Let , be a symmetric, positive-definite inner product on V , so that ω is representedby an invertible skew-adjoint operator K : V → V ; i.e. ω(x, y) = Kx, y . 18
  19. 19. √The operator K admits a polar decomposition K = RJ, where R = KK t is positive-definite symmetric, J = R−1 K is orthogonal, and RJ = JR. From the skew-symmetry of Kit follows that J t = −J, and so J 2 = −JJ t = − id; i.e., J is a complex structure on V . To see that J is ω-compatible, first note that ω(Jx, Jy) = KJx, Jy = JKx, Jy = Kx, y = ω(x, y).Also, ω(x, Jx) = Kx, Jx = JRx, Jx = Rx, x > 0,since R and , are positive-definite. 2Corollary 3.3 The collection J of ω-compatible complex structures on a symplectic vectorspace (V, ω) is contractible.Proof. The association J → gJ described above defines a continuous map from J into thespace P of symmetric, positive-definite bilinear forms on V . By the uniqueness of the polardecomposition, it follows that the map which assigns to a form , the complex structure Jconstructed in the preceding proof is continuous, and the composition J → P → J of thesemaps equals the identity on J . Since P is contractible, this implies the corollary. 2 If J is ω-compatible, a hermitian structure on V is defined by ·, · = gJ (·, ·) + iω(·, ·).As is easily checked, a linear transformation T ∈ GL(V ) which preserves any two of thestructures ω, gJ , J on V preserves the third and therefore preserves the hermitian structure.In terms of the automorphism groups Sp(V ), GL(V, J), O(V ), and U (V ) of ω, J, g, and ·, · ,this means that the intersection of any two of Sp(V ), GL(V, J), O(V ) equals U (V ). To determine the Lie algebra sp(V) of Sp(V ), consider a 1-parameter family of maps etAassociated to some linear map A : V → V . For any v, w ∈ V , we have d ω(etA v, etA w) = ω(Av, w) + ω(v, Aw), dt t=0and so A ∈ sp(V) if and only if the linear map ω ◦ A : V → V ∗ is self-adjoint. Consequently, ˜dim(V ) = 2k implies dim(sp(V)) = dim(Sp(V)) = k(2k + 1).Distinguished subspacesThe ω-orthogonal to a subspace W of a symplectic vector space (V, ω) is called the symplecticorthogonal to W . From the nondegeneracy of the symplectic form, it follows that W ⊥⊥ = W and dim W ⊥ = dim V − dim W 19
  20. 20. for any subspace W ⊂ V . Also, (A + B)⊥ = A⊥ ∩ B ⊥ and (A ∩ B)⊥ = A⊥ + B ⊥for any pair of subspaces A, B of V . In particular, B ⊥ ⊂ A⊥ whenever A ⊂ B. Note that the symplectic orthogonal W ⊥ might not be an algebraic complement to W .For instance, if dim W = 1, the skew-symmetry of ω implies that W ⊂ W ⊥ . More generally,any subspace contained in its orthogonal will be called isotropic. Dually, we note that ifcodim W = 1, then W ⊥ is 1-dimensional, hence isotropic, and W ⊥ ⊂ W ⊥⊥ = W . Ingeneral, spaces W satisfying the condition W ⊥ ⊂ W are called coisotropic or involutive.Finally, if W is self-orthogonal, i.e. W ⊥ = W , then the dimension relation above implies that 1dim W = 2 dim V . Any self-orthogonal subspace is simultaneously isotropic and coisotropic,and is called lagrangian. According to these definitions, a subspace W ⊂ V is isotropic if the restriction of thesymplectic form to W is identically zero. At the other extreme, the restriction of ω to certainsubspaces Z ⊂ V may again be nondegenerate; this is equivalent to saying that Z ∩Z ⊥ = {0}or Z + Z ⊥ = V . Such subspaces are called symplectic.Example 3.4 In E ⊕ E ∗ with its usual symplectic structure, both E and E ∗ are lagrangiansubspaces. It also follows from the definition of this structure that the graph of a linear mapB : E → E ∗ is a lagrangian subspace of E ⊕ E ∗ if and only if B is self-adjoint. If (V, ω) is a symplectic vector space, we denote by V ⊕V the vector space V ⊕V equippedwith the symplectic structure ω ⊕ −ω. If T : V → V is a linear symplectic map, then thegraph of T is a lagrangian subspace of V ⊕ V . The kernel of a nonzero covector α ∈ V ∗ is a codimension-1 coisotropic subspace ker α ofV whose symplectic orthogonal (ker α)⊥ is the distinguished 1-dimensional subspace of ker αspanned by ω −1 (α). ˜Example 3.5 Suppose that (V, ω) is a 2n-dimensional symplectic vector space and W ⊂ Vis any isotropic subspace with dim(W ) < n. Since 2n = dim(W ) + dim(W ⊥ ), there exists anonzero vector w ∈ W ⊥ W . The subspace W of V spanned by W ∪{w} is then isotropic anddim(W ) = dim(W )+1. From this observation it follows that for every isotropic subspace Wof a (finite-dimensional) symplectic vector space V which is not lagrangian, there exists anisotropic subspace W of V which properly contains W . Beginning with any 1-dimensionalsubspace of V , we can apply this remark inductively to conclude that every finite-dimensionalsymplectic vector space contains a lagrangian subspace. Various subspaces of a symplectic vector space are related as follows.Lemma 3.6 If L is a lagrangian subspace of a symplectic vector space V , and A ⊂ V is anarbitrary subspace, then: 20
  21. 21. 1. L ⊂ A if and only if A⊥ ⊂ L. 2. L is transverse to A if and only if L ∩ A⊥ = {0}.Proof. Statement (1) follows from the properties of the operation ⊥ and the equation L = L⊥ .Similarly, L + A = V if and only if (L + A)⊥ = L ∩ A⊥ = {0}, proving statement (2). 2Example 3.7 Note that statement (1) of Lemma 3.6 implies that if L ⊂ A, then A⊥ ⊂ A,and so A is a coisotropic subspace. Conversely, if A is coisotropic, then A⊥ is isotropic,and Example 3.5 implies that there is a lagrangian subspace L with A⊥ ⊂ L. Passing toorthogonals, we have L ⊂ A. Thus, a subspace C ⊂ V is coisotropic if and only if it containsa lagrangian subspace. Suppose that V is a symplectic vector space with an isotropic subspace I and a lagrangiansubspace L such that I ∩ L = 0. If W ⊂ L is any complementary subspace to I ⊥ ∩ L, thenI + L ⊂ W + I ⊥ , and so W ⊥ ∩ I ⊂ I ⊥ ∩ L. Thus, W ⊥ ∩ I ⊂ I ∩ L = 0. Since I ⊥ ∩ W = 0by our choice of W , it follows that I + W is a symplectic subspace of V . A pair L, L of transverse lagrangian subspaces of V is said to define a lagrangiansplitting of V . In this case, the map ω defines an isomorphism L ˜ L∗ , which in turn ∗gives rise to a linear symplectomorphism between V and L ⊕ L equipped with its canonicalsymplectic structure (see Example 3.1). If J is a ω-compatible complex structure on V andL ⊂ V a lagrangian subspace, then L, JL is a lagrangian splitting. By Example 3.5, everysymplectic vector space contains a lagrangian subspace, and since every n-dimensional vectorspace is isomorphic to Rn , the preceding remarks prove the following linear “normal form”result:Theorem 3.8 Every 2n-dimensional symplectic vector space is linearly symplectomorphicto (R2n , ωn ).Theorem 3.2 also implies the following useful result.Lemma 3.9 Suppose that V is a symplectic vector space with a ω-compatible complex struc-ture J and let Tε : V → V be given by Tε (x) = x + εJx. 1. If L, L are any lagrangian subspaces of V , then Lε = Tε (L) is a lagrangian subspace transverse to L for small ε > 0. 2. For any two lagrangian subspaces L, L of V , there is a lagrangian subspace L trans- verse to both L and L.Proof. It is easy to check that Tε is a conformal linear symplectic map, i.e. an isomorphismof V satisfying ω(Tε x, Tε y) = (1 + ε2 ) ω(x, y). Thus, Lε is a lagrangian subspace for allε > 0. Using the inner-product gJ on V induced by J, we can choose orthonormal bases 21
  22. 22. {vi }, {wi } of L and L, respectively, so that for i = 1, · · · , k, the vectors vi = wi span L ∩ L .Then {wi + εJwi } form a basis of Lε , and L , Lε are transverse precisely when the matrixM = {ω(vi , wj + εJwj )} = {ω(vi , wj ) + ε gJ (vi , wj )} is nonsingular. Our choice of basesimplies that ε · id 0 M= 0 A+ε·Bwhere A = {ω(vi , wj )}n n i,j=k+1 and B is some (n−k)×(n−k) matrix. Setting I = span{vi }i=k+1and W = span{wi }n i=k+1 , we can apply Example 3.7 to conclude that A is nonsingular, andassertion (1) follows. To prove (2), observe that for small ε > 0, the lagrangian subspace Lε is transverse toL, L by (1). 2In fact, the statement of preceding lemma can be improved as follows. Let {Li } be a count-able family of lagrangian subspaces, and let Ai , Bi be the matrices obtained with respect to Las in the proof above. For each i, the function t → det(Ai + tBi ) is a nonzero polynomial andtherefore has finitely many zeros. Consequently, the lagrangian subspace Tt (L) is transverseto all Li for almost every t ∈ R.The lagrangian grassmannianThe collection of all unoriented lagrangian subspaces of a 2n-dimensional symplectic vectorspace V is called the lagrangian grassmannian L(V ) of V . A natural action of the groupSp(V ) on L(V ), denoted  : Sp(V ) × L(V ) → L(V ) is defined by (T, L) = L (T ) = T (L).Lemma 3.10 The unitary group associated to an ω-compatible complex structure J on Vacts transitively on L(V ).Proof. For arbitrary L1 , L2 ∈ L(V ), an orthogonal transformation L1 → L2 induces asymplectic transformation L1 ⊕ L∗ → L2 ⊕ L∗ in the manner of Example 3.1, which in turn 1 2gives rise to a unitary transformation L1 ⊕ JL1 → L2 ⊕ JL2 mapping L1 onto L2 . 2The stabilizer of L ∈ L(V ) under the U (V )-action is evidently the orthogonal subgroup ofGl(L) defined with respect to the inner-product and splitting L ⊕ JL of V induced by J (seeExample 3.1). Thus, a (non-canonical) identification of the lagrangian grassmannian withthe homogeneous space U (n)/O(n) is obtained from the map  L U (V ) → L(V ). det2The choice of J also defines a complex determinant U (V ) → S 1 , which induces a fibration JL(V ) → S 1 with 1-connected fiber SU (n)/SO(n), giving an isomorphism of fundamentalgroups π1 (L(V )) π1 (S 1 ) Z. 22
  23. 23. This isomorphism does not depend on the choices of J and L made above. Independence of Jfollows from the fact that J is connected (Corollary 3.3). On the other hand, connectednessof the unitary group together with Lemma 3.10 gives independence of L. Passing to homology and dualizing, we obtain a natural homomorphism H 1 (S 1 ; Z) → H 1 (L(V ); Z).The image of the canonical generator of H 1 (S 1 ; Z) under this map is called the universalMaslov class, µV . The result of the following example will be useful when we extend ourdiscussion of the Maslov class from vector spaces to vector bundles.Example 3.11 If (V, ω) is any symplectic vector space with ω-compatible complex structureJ and lagrangian subspace L, then a check of the preceding definitions shows that mL ((T, L )) = mL (L ) · det2 (T ) Jfor any T ∈ U (V ) and L ∈ L(V ). (Recall that  : Sp(V )×L(V ) → L(V ) denotes the naturalaction of Sp(V ) on L(V )). Now consider any topological space M . If f1 , f2 : M → L(V ) are continuous maps, then ∗ ∗the definition of the universal Maslov class shows that (f1 − f2 )µV equals the pull-back of 1 1the canonical generator of H (S ; R) by the map (mL ◦ f1 )(mL ◦ f2 )−1 .(Here we use the fact that when S 1 is identified with the unit complex numbers, the mul-tiplication map S 1 × S 1 → S 1 induces the diagonal map H 1 (S 1 ) → H 1 (S 1 ) ⊕ H 1 (S 1 )H 1 (S 1 × S 1 ) on cohomology). If T : M → Sp(V ) is any map, we set (T · fi ) = (T, fi ).Since Sp(V ) deformation retracts onto U (V ), it follows that T is homotopic to a mapT : M → U (V ), and so ((T · f1 )∗ − (T · f2 )∗ )µV is obtained via pull-back by (mL ◦ (T · f1 ))(mL ◦ (T · f2 ))−1 .From the first paragraph, it follows that this product equals (mL ◦f1 )(mL ◦f2 )−1 , from whichwe conclude that (f1 − f2 )µV = ((T · f1 )∗ − (T · f2 )∗ )µV . ∗ ∗Symplectic manifoldsTo motivate the definition of a symplectic manifold, we first recall some features of the ndifferential form −dαn = ωn = j=1 dqj ∧ dpj which appeared in our earlier discussion. nFirst, we note that j=1 dqj ∧ dpj defines a linear symplectic structure on the tangent spaceof R2n at each point. In fact: ∂ ∂ ∂ ∂ ∂ ∂ ωn , = δjk ωn , =0 ωn , =0 ∂qj ∂pk ∂qj ∂qk ∂pj ∂pk 23
  24. 24. and so ∂ ∂ ωn ˜ = dpj ωn ˜ = −dqj , ∂qj ∂pjfrom which it is clear that ωn is invertible. ˜ Next, we recall that the hamiltonian vector field associated via Hamilton’s equations toH : R2n → R satisfies XH ωn = dH,or in other words, ˜ −1 XH = ωn (dH),so we see that the symplectic form ωn is all that we need to obtain XH from H. Thisdescription of the hamiltonian vector field leads immediately to the following two invarianceresults. First note that by the skew-symmetry of ωn , LXH H = XH · H = ωn (XH , XH ) = 0,implying that XH is tangent to the level sets of H. This again reflects the fact that the flowof XH preserves energy. Since ωn is closed, we also have by Cartan’s formula (see [1]) LXH ωn = d(XH ωn ) + XH dωn = d2 H = 0.This equation implies that the flow of XH preserves the form ωn and therefore generalizes ourearlier remark that the hamiltonian vector field associated to the 1-dimensional harmonicoscillator is divergence-free. We now see what is needed to do hamiltonian mechanics on manifolds. A 2-form ω on amanifold P is a smooth family of bilinear forms on the tangent spaces of P . By assuming thateach of these bilinear forms is nondegenerate, we guarantee that the equation XH = ω −1 (dH) ˜defines a hamiltonian vector field uniquely for any H. Computing the Lie derivative of Hwith respect to XH LXH H = XH · H = ω (XH )(XH ) = ω(XH , XH ) = 0, ˜we see that the conservation of energy follows from the skew-symmetry of the form ω. Finally, invariance of ω under the hamiltonian flow is satisfied if LXH ω = d(XH ω) + XH dω = 0.Here, the term d(XH ω) = d2 H is automatically zero; to guarantee the vanishing of thesecond term, we impose the condition that ω be closed. Thus we make the following definition:Definition 3.12 A symplectic structure on a manifold P is a closed, nondegenerate2-form ω on P . 24
  25. 25. The condition that ω be nondegenerate means that ω defines an isomorphism of vector ˜bundles T P → T ∗ P , or equivalently, that the top exterior power of ω is a volume form onP , or finally, that ω defines a linear symplectic structure on each tangent space of P . An immediate example of a symplectic manifold is furnished by R2n with its standardstructure ωn = n dqj ∧ dpj (a differential form with constant coefficients and not just j=1a single bilinear form). Darboux’s theorem (Section 4.3) will tell us that this is the localmodel for the general case. In the next section, we will see that the cotangent bundle of anysmooth manifold carries a natural symplectic structure. Generalizing our earlier discussion of distinguished subspaces of a symplectic vector space,we call a submanifold C ⊂ P (co-)isotropic provided that each tangent space Tp C of C is a(co-)isotropic subspace of Tp P . When C is coisotropic, the subspaces Cp = (Tp C)⊥ comprisea subbundle (T C)⊥ of T C known as the characteristic distribution of C. It is integrablebecause ω is closed. Of particular interest in our discussion will be lagrangian submanifoldsof P , which are (co-)isotropic submanifolds of dimension 1 dim(P ). More generally, if L is a 2smooth manifold of dimension 1 dim(P ) and ι : L → P is an immersion such that ι∗ ω = 0, 2we will call the pair (L, ι) a lagrangian immersion.Example 3.13 If C ⊂ P is a hypersurface, then C is a coisotropic submanifold. A simplecheck of definitions shows that if H : P → R is a smooth function having C as a regular levelset, then the hamiltonian vector field XH is tangent to the characteristic foliation of C. If (L, ι) is a lagrangian immersion whose image is contained in C, then Lemma 3.6 impliesthat for each p ∈ L, the characteristic subspace Cι(p) ⊂ Tι(p) C is contained in ι∗ Tp L, and thusXH induces a smooth, nonsingular vector field XH,ι on L. In view of the remarks above, thisassertion generalizes the Hamilton-Jacobi theorem (see the end of Section 2.1) to arbitrarysymplectic manifolds and lagrangian immersions. New symplectic manifolds can be manufactured from known examples by dualizing and bytaking products. The symplectic dual of a manifold (P, ω) consists of the same underlyingmanifold endowed with the symplectic structure −ω. Evidently P and its dual P share thesame (co-)isotropic submanifolds. Given two symplectic manifolds (P1 , ω1 ) and (P2 , ω2 ), theirproduct P1 ×P2 admits a symplectic structure given by the sum ω1 ⊕ω2 . More explicitly, thisform is the sum of the pull-backs of ω1 and ω2 by the projections of P1 × P2 to P1 and P2 . Asis easily verified, the product of (co-)isotropic submanifolds of P1 and P2 is a (co-)isotropicsubmanifold of P1 × P2 . A symplectomorphism from (P1 , ω1 ) to (P2 , ω2 ) is a smooth diffeomorphism f : P1 →P2 compatible with the symplectic structures: f ∗ ω2 = ω1 . A useful connection among duals,products, and symplectomorphisms is provided by the following lemma.Lemma 3.14 A diffeomorphism f : P1 → P2 between symplectic manifolds is a symplecto-morphism if and only if its graph is a lagrangian submanifold of the product P2 × P 1 .The collection Aut(P, ω) of symplectomorphisms of P becomes an infinite-dimensional Liegroup when endowed with the C ∞ topology (see [49]). In this case, the corresponding Lie 25
  26. 26. algebra is the space χ(P, ω) of smooth vector fields X on P satisfying LX ω = 0.Since LX ω = d(X ω), the association X → X ω defines an isomorphism between χ(P, ω)and the space of closed 1-forms on P ; those X which map to exact 1-forms are simply thehamiltonian vector fields on P . The elements of χ(P, ω) are called locally hamiltonianvector fields or symplectic vector fields.Example 3.15 A linear symplectic form ω on a vector space V induces a symplectic struc-ture (also denoted ω) on V via the canonical identification of T V with V ×V . The symplecticgroup Sp(V ) then embeds naturally in Aut(V, ω), and the Lie algebra sp(V) identifies withthe subalgebra of χ(V, ω) consisting of vector fields of the form X(v) = Avfor some A ∈ sp(V). Note that these are precisely the hamiltonian vector fields of thehomogeneous quadratic polynomials on V , i.e. functions satisfying Q(tv) = t2 Q(v) for allreal t. Consequently, sp(V) is canonically identified with the space of such functions via thecorrespondence 1 A ↔ QA (v) = ω(Av, v). 2Symplectic vector bundlesSince a symplectic form on a 2n-manifold P defines a smooth family of linear symplecticforms on the fibers of T P , the frame bundle of P can be reduced to a principal Sp(n) bundleover P . More generally, any vector bundle E → B with this structure is called a symplecticvector bundle. Two symplectic vector bundles E, F are said to be symplectomorphic ifthere exists a vector bundle isomorphism E → F which preserves their symplectic structures.Example 3.16 If F → B is any vector bundle, then the sum F ⊕ F ∗ carries a naturalsymplectic vector bundle structure, defined in analogy with Example 3.1. With the aid of an arbitrary riemannian metric, the proof of Theorem 3.2 can be gener-alized by a fiberwise construction as follows.Theorem 3.17 Every symplectic vector bundle admits a compatible complex vector bundlestructure.Example 3.18 Despite Theorem 3.17, there exist examples of symplectic manifolds whichare not complex (the almost complex structure coming from the theorem cannot be madeintegrable), and of complex manifolds which are not symplectic. (See [27] and the numerousearlier references cited therein.) Note, however, that the K¨hler form of any K¨hler manifold a ais a symplectic form. 26
  27. 27. A lagrangian subbundle of a symplectic vector bundle E is a subbundle L ⊂ E suchthat Lx is a lagrangian subspace of Ex for all x ∈ B. If E admits a lagrangian subbundle,then E is symplectomorphic to L ⊕ L∗ , and the frame bundle of E admits a further reductionto a principal GL(n) bundle over B (compare Example 3.1).Example 3.19 If L is a lagrangian submanifold of a symplectic manifold P , then the re-stricted tangent bundle TL P is a symplectic vector bundle over L, and T L ⊂ TL P is alagrangian subbundle. Also note that if C ⊂ P is any submanifold such that T C contains alagrangian subbundle of TC P , then C is coisotropic (see Lemma 3.6). In general, the automorphism group of a symplectic vector bundle E does not act tran-sitively on the lagrangian subbundles of E. Nevertheless, a pair of transverse lagrangiansubbundles can be related as follows.Theorem 3.20 Let E → B be a symplectic vector bundle and suppose that L, L are la-grangian subbundles such that Lx is transverse to Lx for each x ∈ B. Then there exists acompatible complex structure J on E satisfying JL = L .Proof. Let J0 be any compatible complex structure on E. Since L and J0 L are bothtransverse to L, we can find a symplectomorphism T : L ⊕ L → L ⊕ J0 L which preservesthe subbundle L and maps L to J0 L. A simple check of the definition then shows thatJ = T −1 J0 T is a compatible complex structure on E which satisfies JL = L . 2Example 3.21 If E is a symplectic vector bundle over M , then any pair L, L of lagrangiansubbundles of E define a cohomology class µ(L, L ) ∈ H 1 (M ; Z) as follows. Assuming first that E admits a symplectic trivialization f : E → M × V for somesymplectic vector space V , we denote by fL , fL : M → L(V ) the maps induced by thelagrangian subbundles f (L), f (L ) of M × V . Then ∗ ∗ µ(L, L ) = (fL − fL ) µV ,where µV ∈ H 1 (L(V ); Z) is the universal Maslov class. From Example 3.11 it follows thatthis class is independent of the choice of trivialization f . For nontrivial E, we note that since the symplectic group Sp(V ) is connected, it followsthat for any loop γ : S 1 → M , the pull-back bundle γ ∗ E is trivial. Thus µ(L, L ) is well-defined by the requirement that for every smooth loop γ in M , γ ∗ µ(L, L ) = µ(γ ∗ L, γ ∗ L ). 27
  28. 28. Example 3.22 As a particular case of Example 3.21, we consider the symplectic manifoldR2n T ∗ Rn with its standard symplectic structure. Then the tangent bundle T (R2n ) isa symplectic vector bundle over R2n with a natural “vertical” lagrangian subbundle V Rndefined as the kernel of π∗ , where π : T ∗ Rn → Rn is the natural projection. If ι : L → R2n is a lagrangian immersion, then the symplectic vector bundle ι∗ T (T ∗ (Rn ))has two lagrangian subbundles, the image L1 of ι∗ : T L → ι∗ T (T ∗ (Rn )) and L2 = ι∗ V Rn .The class µL,ι = µ(L1 , L2 ) ∈ H 1 (L; Z) is called the Maslov class of (L, ι). A check of these definitions shows that the Maslov class of (L, ι) equals the pull-back ofthe universal Maslov class µn ∈ H 1 (L(R2n ); Z) by the Gauss map G : L → L(R2n ) definedby G(p) = ι∗ Tp L ⊂ R2n . (See [3] for an interpretation of the Maslov class of a loop γ inL as an intersection index of the loop G ◦ γ with a singular subvariety in the lagrangiangrassmannian).3.2 Cotangent bundlesThe cotangent bundle T ∗ M of any smooth manifold M is equipped with a natural 1-form,known as the Liouville form, defined by the formula αM ((x, b))(v) = b(π∗ v),where π : T ∗ M → M is the canonical projection. In local coordinates (x1 , · · · , xn ) on M andcorresponding coordinates (q1 , · · · , qn , p1 , · · · , pn ) on T ∗ M , the equations ∂ qj (x, b) = xj (x) pj (x, b) = b ∂xjimply that n αM = pj dqj . j=1 nThus, −dαM = dqj ∧ dpj in these coordinates, from which it follows that the form j=1ωM = −dαM is a symplectic structure on T ∗ M . Note that if M = Rn , then ωM is just thesymplectic structure ωn on T ∗ Rn R2n discussed previously, and αM = αn .Lagrangian immersions and the Liouville classGiven a lagrangian immersion ι : L → T ∗ M , we set πL = π ◦ ι, where π : T ∗ M → Mis the natural projection. Critical points and critical values of πL are called respectivelysingular points and caustic points of L. Finally, (L, ι) is said to be projectable ifπL is a diffeomorphism. A nice property of the Liouville 1-form is that it can be used toparametrize the set of projectable lagrangian submanifolds. To do this, we use the notationιϕ to denote a 1-form ϕ on M when we want to think of it as a map from M to T ∗ M . 28
  29. 29. Lemma 3.23 Let ϕ ∈ Ω1 (M ). Then ι∗ αM = ϕ. ϕProof. Because ιϕ is a section of T ∗ M , it satisfies π ◦ ιϕ = idM . By the definition of αM , itfollows that for each v ∈ Tp M , ι∗ αM (p)(v) = αM (ιϕ (p))(ιϕ∗ v) = ιϕ (p), π∗ (ιϕ∗ v) = ιϕ (p), v . ϕ 2For this reason, αM is often described as the “tautological” 1-form on T ∗ M . Taking exteriorderivatives on both sides of the equation in Lemma 3.23, we get dϕ = dι∗ αM = ι∗ dαM = −ι∗ ωM . ϕ ϕ ϕFrom this equation we see that the image of ϕ is a lagrangian submanifold of T ∗ M preciselywhen the form ϕ is closed. This provesProposition 3.24 The relation ϕ ↔ (M, ιϕ ) defines a natural bijective correspondence be-tween the the vector space of closed 1-forms on M and the set of projectable lagrangiansubmanifolds of T ∗ M .Generalizing our WKB terminology, we will call S : M → R a phase function for aprojectable lagrangian embedding (L, ι) ⊂ T ∗ M provided that ι(L) = dS(M ). The precedingremarks imply a simple link between phase functions and the Liouville form:Lemma 3.25 If (L, ι) ⊂ T ∗ M is a projectable lagrangian embedding, then S : M → R is aphase function for L if and only if d(S ◦ πL ◦ ι) = ι∗ αM .Thus, L is the image of an exact 1-form on M if and only if the restriction of the Liouvilleform to L is itself exact. This motivates the following definition.Definition 3.26 If L, M are n-manifolds and ι : L → T ∗ M is an immersion such that ι∗ αMis exact, then ι is called an exact lagrangian immersion.If ι : L → T ∗ M is an exact lagrangian immersion, then Lemma 3.25 suggests that theprimitive of ι∗ αM is a sort of generalized phase function for (L, ι) which lives on the manifoldL itself. We will return to this important viewpoint in the next chapter.Example 3.27 A simple application of Stokes’ theorem shows that an embedded circle inthe phase plane cannot be exact, although it is the image of an exact lagrangian immersionof R. A general class of exact lagrangian submanifolds can be identified as follows. Associatedto a smooth submanifold N ⊂ M is the submanifold N ⊥ = {(x, p) ∈ T ∗ M : x ∈ N, Tx N ⊂ ker(p)},known as the conormal bundle to N . From this definition it follows easily that dim T ∗ M =2 dim N ⊥ , while the Liouville form of T ∗ M vanishes on N ⊥ for any N . If F is a smooth foliation of M , then the union of the conormal bundles to the leaves of Fis a smooth submanifold of T ∗ M foliated by lagrangian submanifolds and is thus coisotropic(see Example 3.19). 29
  30. 30. Although many lagrangian immersions ι : L → T ∗ M are not exact, the form ι∗ αM isalways closed, since dι∗ αM = ι∗ ωM = 0. The deRham cohomology class λL,ι ∈ H 1 (L; R)induced by this form will play an important role in the quantization procedures of the nextchapter and is known as the Liouville class of (L, ι).Example 3.28 To generalize the picture described in Example 3.27, we consider a smoothmanifold M , together with a submanifold N ⊂ M and a closed 1-form β on N . Then Nβ = {(x, p) ∈ T ∗ M : x ∈ N p|Tx N = β(x)} ⊥is a lagrangian submanifold of T ∗ M whose Liouville class equals [πN β] ∈ H 1 (Nβ ; R), where ∗ ⊥πN : Nβ → N is here the restriction of the natural projection π : T ∗ M → M . ⊥Fiber-preserving symplectomorphismsOn each fiber of the projection π : T ∗ M → M , the pull-back of αM vanishes, so the fibers arelagrangian submanifolds. Thus, the vertical bundle V M = ker π∗ is a lagrangian subbundleof T (T ∗ M ). Since αM vanishes on the zero section ZM ⊂ T ∗ M , it follows that ZM islagrangian as well, and the subbundles T ZM and V M define a canonical lagrangian splittingof T (T ∗ M ) over ZM . A 1-form β on M defines a diffeomorphism fβ of T ∗ M by fiber-wise affine translation fβ (x, p) = (x, p + β(x)).It is easy to see that this map satisfies fβ αM = αM + π ∗ β, ∗so fβ is a symplectomorphism of T ∗ M if and only if β is closed.Theorem 3.29 If a symplectomorphism f : T ∗ M → T ∗ M preserves each fiber of the pro-jection π : T ∗ M → M , then f = fβ for a closed 1-form β on M .Proof. Fix a point (x0 , p0 ) ∈ T ∗ M and let ψ be a closed 1-form on M such that ψ(x0 ) =(x0 , p0 ). Since f is symplectic, the form µ = f ◦ ψ is also closed, and thus the map −1 h = fµ ◦ f ◦ fψis a symplectomorphism of T ∗ M which preserves fibers and fixes the zero section ZM ⊂ T ∗ M .Moreover, since the derivative Dh preserves the lagrangian splitting of T (T ∗ M ) along ZMand equals the identity on T ZM , we can conclude from Example 3.1 that Dh is the identityat all points of ZM . Consequently, the fiber-derivative of f at the arbitrary point (x0 , p0 )equals the identity, so f is a translation on each fiber. Defining β(x) = f (x, 0), we havef = fβ . 30
  31. 31. 2 ω −1If β is a closed 1-form on M , then the flow ft of the vector field Xβ = −˜ M (π ∗ β) issymplectic; since V M ⊂ ker π ∗ β, the Hamilton-Jacobi theorem implies furthermore that theflow ft satisfies the hypotheses of Theorem 3.29.Corollary 3.30 For any closed 1-form β on M , the time-1 map f = f1 of the flow of Xβequals fβ .Proof. By Theorem 3.29, the assertion will follow provided that we can show that f ∗ αM =αM + π ∗ β. To this end, note that the definition of the Lie derivative shows that f satisfies 1 1 ∗ d ∗ f αM = αM + (f αM ) dt = αM + ft∗ (LXβ αM ) dt. 0 dt t 0By Cartan’s formula for the Lie derivative, we have LXβ αM = d(Xβ αM ) − Xβ ωM = π ∗ β,the latter equality following from the fact that Xβ ⊂ V M ⊂ ker αM and dαM = −ωM .Another application of Cartan’s formula, combined with the assumption that β is closedshows that LXβ π ∗ β = 0,and so ft∗ π ∗ β = π ∗ β for all t. Inserting these computations into the expression for f ∗ αMabove, we obtain f ∗ αM = αM + π ∗ β. 2Using Theorem 3.29, we can furthermore classify all fiber-preserving symplectomorphismsfrom T ∗ M to T ∗ N .Corollary 3.31 Any fiber-preserving symplectomorphism F : T ∗ M → T ∗ N can be realizedas the composition of a fiber-translation in T ∗ M with the cotangent lift of a diffeomorphismN → M.Proof. By composing F with a fiber-translation in T ∗ M we may assume that F maps thezero section of T ∗ M to that of T ∗ N . The restriction of F −1 to the zero sections then inducesa diffeomorphism f : N → M such that the composition F ◦ (f −1 )∗ is a fiber-preservingsymplectomorphism of T ∗ N which fixes the zero section. From the preceding theorem, weconclude that F = f ∗ . 2 31
  32. 32. The Schwartz transformIf M, N are smooth manifolds, then the map SM,N : T ∗ M × T ∗ N → T ∗ (M × N ) defined inlocal coordinates by ((x, ξ), (y, η)) → (x, y, −ξ, η)is a symplectomorphism which we will call the Schwartz transform.5 An elementary, butfundamental property of this mapping can be described as follows.Proposition 3.32 If M, N are smooth manifolds, then the Schwartz transform SM,N satis-fies (SM,N )∗ αM ×N = αM ⊕ −αN .In particular, SM,N induces a diffeomorphism of zero sections ZM × ZN ZM ×Nand an isomorphism of vertical bundles VM ⊕VN V (M × N ). Using the Schwartz transform, we associate to any symplectomorphism F : T ∗ M → T ∗ Nthe lagrangian embedding ιF : T ∗ M → T ∗ (M × N ) defined as the composition of SM,N withthe graph ΓF : T ∗ M → T ∗ M × T ∗ N .Example 3.33 By Corollary 3.31, a fiber-preserving symplectomorphism F : T ∗ M → T ∗ Nequals the composition of fiber-wise translation by a closed 1-form β on M with the cotangentlift of a diffeomorphism g : N → M . A computation shows that if Γ ⊂ M × N is the graphof g and p : Γ → M is the natural projection, then the image of the composition of thelagrangian embedding (T ∗ M, ιF ) with the Schwartz transform SM,N equals the submanifoldΓ⊥∗ β ⊂ T ∗ (M × N ) defined in Example 3.28. In particular, if F is the cotangent lift of g, pthen the image of (T ∗ M, ιF ) equals the conormal bundle of Γ. Finally, we note that multiplying the cotangent vectors in T ∗ M by −1 defines a sym-plectomorphism T ∗ M → T ∗ M which can be combined with the Schwartz transform SM,Nto arrive at the usual symplectomorphism T ∗ M × T ∗ N T ∗ (M × N ). Thus in the specialcase of cotangent bundles, dualizing and taking products leads to nothing new.3.3 Mechanics on manifoldsWith the techniques of symplectic geometry at our disposal, we are ready to extend ourdiscussion of mechanics to more general configuration spaces. Our description begins with acomparison of the classical and quantum viewpoints, largely parallelling the earlier materialon the 1-dimensional harmonic oscillator given in the introduction. We then turn to thesemi-classical approximation and its geometric counterpart in this new context, setting thestage for the quantization problem in the next chapter. 5 The name comes from the relation of this construction to the Schwartz kernels of operators (see Sec-tion 6.2). 32
  33. 33. The classical pictureThe hamiltonian description of classical motions in a configuration space M begins with theclassical phase space T ∗ M . A riemannian metric g = (gij ) on M induces an inner producton the fibers of the cotangent bundle T ∗ M , and a “kinetic energy” function which in localcoordinates (q, p) is given by 1 kM (q, p) = g ij (q) pi pj , 2 i,jwhere g ij is the inverse matrix to gij . Regular level sets of kM are sphere bundles over M .The hamiltonian flow associated to kM is called the co-geodesic flow due to its relationwith the riemannian structure of M described in the following theorem.Theorem 3.34 If M is a riemannian manifold, then integral curves of the co-geodesic flowproject via π to geodesics on M .A proof of this theorem can be given in local coordinates by using Hamilton’s equations ∂H ∂H 1 ∂g uv qi = ˙ = g ij pj pi = − ˙ =− pu pv ∂pi j ∂qi 2 u,v ∂qito derive the geodesic equation qk + ¨ Γk qi q˙j = 0. ij ˙ i,jFor details, see [36]. In physical terms, Theorem 3.34 states that a free particle on a manifold must move alonga geodesic. A smooth, real-valued potential V : M → R induces the hamiltonian function H(q, p) = kM (q, p) + V (q)on T ∗ M . Integral curves of the hamiltonian flow of H then project to classical trajectoriesof a particle on M subject to the potential V .The quantum mechanical pictureFor the time being, we will assume that the Schr¨dinger operator on a riemannian manifold oM with potential function V is defined in analogy with the flat case of Rn with its standardmetric. That is, we first define the operator on the function space C ∞ (M ) by 2 ˆ H=− ∆ + mV , 2m ˆwhere ∆ denotes the Laplace-Beltrami operator. As before, H induces a (densely defined) ˆ on the intrinsic Hilbert space HM of M by the equationoperator H ˆ ˆ H(a|dx|1/2 ) = (Ha)|dx|1/2 , 33
  34. 34. where |dx| is the natural density associated to the metric on M , and the time-independentSchr¨dinger equation on M assumes the familiar form o ˆ (H − E)ϕ = 0.The advantage of this viewpoint is that both the classical state space T ∗ M and the quantumstate space HM are objects intrinsically associated to the underlying differential manifoldM . The dynamics on both objects are determined by the choice of metric on M .The semi-classical approximationThe basic WKB technique for constructing semi-classical solutions to the Schr¨dinger equa- otion on M proceeds as in Section 2.2. Specifically, a half-density of the form eiS/h a is a ˆsecond-order approximate solution of the eigenvector problem (H − E) ϕ = 0 provided thatthe phase function S : M → R satisfies the Hamilton-Jacobi equation H ◦ ιdS = E,and the half-density a satisfies the homogeneous transport equation, which assumes thecoordinate-free form a∆S + 2L S a = 0. We can formulate this construction abstractly by considering first a projectable, exactlagrangian embedding ι : L → T ∗ M . By definition, this means that πL = π ◦ ι is a diffeo-morphism, where π : T ∗ M → M is the natural projection, and Lemma 3.25 implies that for −1any primitive φ : L → R of ι∗ αM , the composition S = φ ◦ πL is a phase function for (L, ι). Now if H : T ∗ M → R is any smooth function, then (L, ι) satisfies the Hamilton-Jacobiequation provided that E is a regular value of H and H ◦ ι = E.In this case, the embedding ι and hamiltonian vector field XH of H induce a nonsingularvector field XH,ι on L (see Example 3.13). If a is a half-density on L, then the requirement −1that (πL )∗ a satisfy the homogeneous transport equation on M becomes LXH,ι a = 0.If these conditions are satisfied, then the half-density eiφ/ a on L can be quantized (i.e. −1pulled-back) to yield a second-order approximate solution (πL )∗ eiφ/ a to the Schr¨dinger oequation on M , as above. This interpretation of the WKB approximation leads us to consider a semi-classical stateas a quadruple (L, ι, φ, a) comprised of a projectable, exact lagrangian embedding ι : L